Exhibitor portal adoption measured in completed tasks per invited exhibitor

The slide says portal adoption is 98 per cent. It means 392 of 400 exhibiting companies have logged in at least once. Nobody in the room believes the floor is 98 per cent ready, and three weeks later the operations team is still chasing certificates of insurance from a third of the book.

Exhibitor portal adoption measured as accounts created has this property at every show. Logging in is the one thing an exhibitor does automatically, because the contract confirmation email contains the link and they click it to see what they bought. It is a measure of whether your invitation email works.

A better definition takes about a day to implement and produces a number that moves for reasons you can act on. Adoption is completed required tasks divided by required tasks, summed across invited exhibitors, measured weekly.

Why is account creation the wrong number?

Because it saturates, and a metric that saturates cannot rank anything.

At the 400 exhibitor show above, the login count reaches the high nineties within a fortnight of the invitations going out and then sits there for four months. The one number your operations team most wants, which is whether onboarding is ahead or behind last edition, cannot be read off a series that is flat at 98.

It also inverts the risk. The eight companies that never logged in are visible and get chased. The 130 companies that logged in once in April, looked at the floor plan and never came back are invisible, and they are the ones who will arrive at move in without insurance.

Nielsen Norman Group (2016) reported that forms following usability guidelines produced 78 per cent one-try submissions against 42 per cent for forms that did not, which is a reminder that completion is a property of the screen as much as of the person. Applied here, the point is that adoption has to be measured at the level where the work happens, and the work happens task by task.

Define the denominator before you count anything

The denominator is where this measure is usually broken, and fixing it is most of the job.

Take the required task list. On a typical B2B show it runs to something like nine items: company profile, product categories, show directory listing, booth staff badges, booth services order, certificate of insurance, W-9, rigging and hanging sign plan, and a booth layout drawing.

Two of those nine do not apply to everybody. Only exhibitors with a hanging sign file a rigging plan. Only exhibitors above a certain size or with an island booth file a layout drawing. If you divide by nine for every exhibitor, you are counting a 10 by 10 inline as incomplete for a task it can never do, and the show's adoption figure falls every time you sell more small booths.

So the denominator is per exhibitor. For each company, count the tasks the show actually requires of them given their booth type, their contractor arrangement and their country of registration. Sum those counts across the book. That total is the denominator, and it changes as booths get sold and upgraded, which means it has to be recomputed each week rather than fixed at launch.

Do the same discipline on the numerator. A task counts as complete once the show has accepted it. Pressing submit only starts the check, and an uploaded certificate that fails the additional insured test leaves an incomplete task with a document attached. Counting that as done is how a portal reports 80 per cent adoption into a move in where a quarter of the hall has no valid cover.

The weekly measure, worked through

Figures below are invented so the arithmetic can be followed on the page.

At 30 days before move in, the 400 exhibitor show has these completion counts across the nine required tasks: company profile 388, product categories 361, booth staff badges 342, show directory listing 296, booth services order 250, certificate of insurance 232, W-9 190, rigging and hanging sign plan 61, booth layout drawing 40. Those sum to 2,160 completed tasks.

The naive denominator is 9 times 400, or 3,600. That gives 2,160 over 3,600, which is 60.0 per cent, and a mean of 5.4 completed tasks per exhibitor.

The correct denominator is smaller. Seven tasks apply to all 400 companies, which is 2,800. The rigging plan applies to the 78 companies with a hanging sign. The layout drawing applies to the 96 companies with an island or peninsula booth. Total required tasks is 2,800 plus 78 plus 96, or 2,974.

Adoption is then 2,160 over 2,974, which is 72.6 per cent.

The denominator choice moved the headline by 12.6 points, and only one of the two numbers can be compared against last edition. Whichever you pick, write the rule down and never change it mid-season, because a show that quietly fixes its denominator in week nine has destroyed its own trend line.

What does a 72 per cent adoption figure actually tell you?

On its own it says very little, which is why the task level split matters more than the headline.

The same file says company profile is at 97.0 per cent and booth layout drawing is at 41.7 per cent of the 96 companies that owe one. Certificate of insurance is at 58.0 per cent. The single aggregate averages a task nearly everyone has done with a task most people have not, and the operations consequence of those two being incomplete could hardly be more different.

So publish two rates. The flat rate across all required tasks, which is 72.6 per cent, and the blocking rate across only the tasks that stop a booth going up. If the blocking set is booth staff badges, certificate of insurance, rigging plan and layout drawing, that is 342 plus 232 plus 61 plus 40, or 675 completed, over 800 plus 78 plus 96, or 974 required. The blocking rate is 69.3 per cent, and it is the number that belongs on the operations slide.

A show reporting 98 per cent logins and 69.3 per cent blocking task completion at 30 days out is describing two different realities, and only one of them is about the floor opening.

Adoption per task tells you where to spend the redesign

Rank the nine tasks by completion rate at a fixed point in the window and the bottom of that list is your work list for next edition.

Nielsen Norman Group (2025) sets out an approach for reducing form effort in three moves: eliminate the questions that are nonessential or irrelevant, automate what the system can fill in, and simplify what remains. Run that against the two worst tasks rather than against the whole portal. If the booth layout drawing sits at 41.7 per cent, ask whether the show needs the drawing from every island, whether it can be prefilled from the floor plan file it already holds, and whether the upload accepts the format exhibitors already have.

Some low completion rates are correct and should be left alone. A task with a genuine third party dependency, such as a certificate issued by a broker, completes late because brokers are slow, and no amount of form design fixes that. Distinguishing a slow task from a broken one needs the shape of the curve over time, which is what completion rates by task type across the onboarding window sets out.

Reporting it so it survives a room

Report the weekly series against the same week of the previous edition, indexed to days before move in. Calendar weeks are useless for this, because show dates move and a 12 May reading means nothing next to a 4 June one.

Publish the denominator next to the rate every time. A single line saying 2,974 required tasks across 400 exhibitors, 2,160 complete, 72.6 per cent, is enough for anyone in the room to check the division, and it stops the argument about what the number counts before the argument starts.

Keep the chase list attached. Of the 400 companies, if 41 have completed one or none of their required tasks, that list is the actual output of the measurement, and it is what a coordinator opens on Monday morning. The portal event log read after the show does something related for exhibitor engagement, and the two draw on the same underlying table for different purposes.

Where this stops

This measure assumes the required task list is right, and on most shows it is a historical accident. Tasks accumulate. Somebody added a question in 2019 after one bad incident, and it has been required ever since, and nobody has asked whether the answer is read. A portal adoption rate computed over a task list padded with dead requirements will look low for a reason that has nothing to do with exhibitor behaviour, and the fix is an audit of the task list rather than a chase campaign.

It also cannot see work that happens outside the portal. An exhibitor who emails their certificate of insurance to the operations coordinator has done the task, and unless someone marks it complete in the system, they appear on the chase list and get chased, which is the fastest way to lose an exhibitor's trust in the portal entirely. Every route into the show has to write back to the same task record, and where inbound exhibitor email piles up tells you how much of the work is currently bypassing the system.

The honest limit is that adoption is a proxy. What operations cares about is whether the hall opens without incident, and the relationship between a task completion rate and that outcome is one every show has to measure for itself over several editions.

Take last edition's portal database, pick the 30 days before move in mark, and compute both numerators and both denominators for a single show. If the two adoption figures differ by more than a few points, the denominator conversation is the one to have this week, before anyone builds a dashboard on top of the wrong one. The rest of the onboarding picture, including compliance and cycle time, reads from the same task records.